Policy networks for European internal security governance: toward a more systematic empirical and normative assessment

European cooperation on internal security is usually seen through the lens of highly salient normative debates, such as the appropriate balance between freedom and security. This paper provides an alternative and long-term perspective on this policy area. The first theoretical part condenses four alternative theoretical perspectives on European internal security governance that relate to the wider relationship between informal and formal governance mechanisms. The second part presents a matching empirical overview of more than 180 European governance networks on police and criminal cooperation, counterterrorism, border security, migration policy, cybersecurity and disaster prevention. Expertise-driven and policy field-specific dynamics appear significant to account for the breadth and variety of policy networks. In contrast, executive empowerment, venue-shopping or deliberate side-stepping of formal decision-making forums do not appear as dominant trends, although there are severe structural shortcomings with regard to the transparency and legal basis of the mapped policy networks.


Introduction
European cooperation on internal security is becoming increasingly complex and more controversial. Particularly, since the parallel crises of large-scale irregular migration and intensified international terrorism between 2015 and 2018, mainly inspired by the Islamic state, European Heads of States and Government have become preoccupied with legitimating or defending the EU's role in this field.
The EU is now supposed to be a "Union that protects," 1 even if member states and national security authorities remain legally responsible for any operational or executive measures in this regard. European citizens are also apparently in favor of more integrated and robust security policies to control the EU's borders and to combat terrorism. 2 At the same time, there is an intense political debate about the legitimacy of EU security policies, especially with regard to the treatment of irregular migrants and asylum seekers both within the Union and in neighboring countries. 3 Across the Atlantic, a politically different 4 but structurally partly convergent 5 confrontation over border security and irregular migration has been stoked by the Trump administration.
However, this paper deliberately takes a step back from these highly salient public debates and seeks to provide a different as well as long-term perspective on the wider landscape of European internal security governance. It includes the currently most pronounced subfields of the fight against international terrorism and border security, but also many other channels for police cooperation, cybersecurity or critical infrastructures and disaster prevention. Furthermore, the geographical contours of this landscape of European internal security cooperation are extremely fuzzy, ranging from sub-groups of European member states in matters of regular police work to some important transatlantic networks in the area of counterterrorism or intelligence.
Based on this very broad empirical understanding, the paper engages with different theoretical perspectives on the evolution of this field, which-well before the current public crisis-has been deeply contested among scholars. On the one hand, the rise of transboundary security problems and shared vulnerability has supported a functionalist narrative, whereby security actors have to "catch up" and forge new flexible ways of international cooperation. As in other policy fields, transnational networking can also help to foster a greater professional expertise and evidence-led policy-making with regard to ever more complex security challenges. On the other hand, the gradual shift of "core state powers" to a supranational level has been problematized with a view to its eventual constitutional end point or settlement. 6 In other words, can the EU develop a fully fledged system of political, executive and judicial competences in the area of internal security that resemble the liberal constitutions of its member states? Even though the Treaty of Lisbon integrated the so-called Area of Freedom, Security and Justice (AFSJ)-or EU internal security policy-into the regular decision-making system of the EU some 10 years ago, the jury is still out. 7 Some critical security scholars have gone further and argued that transnational networks are fueling an excessive reliance on security practices to address social and political challenges 8 or are eroding the classical system of checks-and-balances and democratic security policy-making. 9 The main aim of this paper is to go beyond these largely opposing theoretical camps, which often base their arguments on specific case studies or historical moments, and to probe their plausibility with regard to the wider landscape of European internal security governance. To do so, the article proceeds in two steps. The first theoretical section condenses four alternative theoretical perspectives on European internal security governance. To aid this theoretical synthesis, the article refers back to wider debates in governance as well as EU studies that revolve around the relationship between informal and formal governance mechanisms. This calls for a nuanced understanding of the intersection between formal and hierarchical, on the one hand, and informal and network-like governance, on the other. Different options for a complementary or competing relationship of these general governance types can be related to the more specific debate on the effectiveness and legitimacy of European internal security governance.
The second part of this article seeks to present a broad empirical overview of different governance networks and informal venues that shape EU internal security cooperation, matching with the theoretical elaborations. A database drawn up by the author lists more than 180 entries for governance networks on police and criminal cooperation, counterterrorism, border security, migration policy, cybersecurity and disaster prevention. Even if still provisional, this overview allows to distill some general empirical patterns that can be related to the four condensed theoretical perspectives on EU internal security governance. Expertise-driven and policy field-specific dynamics appear significant to account for the breadth and variety of policy networks. In contrast, executive empowerment, venue-shopping or deliberate side-stepping of formal decision-making forums do not appear as dominant trends, although there are severe structural shortcomings with regard to the transparency and legal basis of the mapped policy networks. In light of these tentative patterns, there is a substantial need for further research to fill in knowledge gaps with regard to available oversight and accountability mechanisms for informal networks for European internal security governance.

Informal and formal governance
The concept of governance is usually used to emphasize soft or rather horizontal coordination mechanisms against hierarchical decision-making and control processes. Yet, there are widely different interpretations of governance. In general, one can distinguish narrow interpretations, which emphasize the contrast between networks and hierarchical regulation, and broad interpretations, in which the whole of market-based, hierarchical as well as network-shaped coordination processes are taken into account. 10 Powerful public actors, such as the EU, may assume an overarching coordination role in practice, whereas a strict reading of the original meaning of governance would not allow for the possibility of central network management. 11 At the same time, there are a variety of issue-specific takes, such as "Security Governance." 12 International Relations as well as EU studies have seen a growing debate on the role of "informality." 13 The central argument is that informality has multifarious and partially highly productive functions and cannot only be understood as a counterpoint to formal decision making. This contrasts with previous research agendas, which focused on the legalization of international regimes 14 or constitutionalization of the EU. 15 In many current contributions, the interplay between official and unofficial coordination mechanisms in transnational space is considered more fruitful.
However, the respective context of the policy field, the type of actors involved and the time dimension for decision-making processes (e.g., in crises) each demand different approaches, while the idea of informality can also be understood in diverse ways. 16 Yet, the following scheme adapted 17 provides some overarching guidance. The complementarity of the objectives of formal and informal governance (vertical axis) as well as the effectiveness of formal decision-making processes (horizontal axis) is often examined. In traditional approaches, the absence of a clear political order in transnational space would be the reason for the emergence of informal governance (in a substitute role, right above). Conversely, a decline in informal governance would be expected when EU integration or other forms of supranational institutionalization take place. The more recent expanded conceptual understanding does not see any direct contrast, but the options of complementary (upper left) and competitive (the lower one) co-existence between formal and informal governance. 16 Bröchler and Grunden [8]. 17 Christiansen and Neuhold, "Informal Politics in the EU," 1201. 10 Mayntz [52]. 11 Grande [32]. 12 Frevel and Schulze [28]. 13 Verdier [77], Kleine [44], Pauwelyn et al. [61], Christiansen and Neuhold [11], Stacey [70]. 14 List and Zangl [49]. 15 Eriksen et al. [21]. This article argues that research on European cooperation in internal security has mainly focused on divergent or competitive dynamics between formal and informal governance. For example, one can refer to the long-standing debate on "venue shopping," 18 which is the search for venues where divergent informal policy results should be possible. Sociological work on professional security networks and fields 19 leads to similar assumptions, namely the circumvention or overwriting of formal policy processes by informal dynamics. This extensive literature cannot be reviewed in detail here. Rather, this chapter seeks to compress different theoretical perspectives that can inform alternative assessments of the complex landscape of European internal security governance. These perspectives do not correspond exactly to the four quadrants of the scheme by Christiansen and Neuhold, but raise parallel questions about the effectiveness and complementarity of governance processes: 1. The first perspective-related to the top right quadrant-assumes the gradual replacement of historically grown informal governance by increasingly formal institutional one. That is, the growth of European integration and institutionbuilding obviates the need for supporting, or trail-blazing, informal governance. Such an integration dynamics is the result of increasing functional constraints, transaction costs of international cooperation and the gradual adaptation of interests-or abandoned sovereignty-of state actors. 20 Variants of EU integration theory, such as neo-functionalism, or IB theories on the growing institutionalization and path-dependence of regimes 21 can be drawn upon for this perspective. 2. The second perspective-related to the bottom-right quadrant-sees informal governance as a means to circumvent formal decision-making structures. 22 This can be motivated by general interests to increase leeway of action for executive actors. 23 Ministries of the Interior and various security authorities thus use transnational governance to achieve policy goals, which are not achievable in national structures. Alternatively, the interests of individual influential actors (e.g. large states) can explain the use of informal and flexible cooperation processes, which sidestep formal institutional rules that stress equality. 24 3. The third perspective-related to the top-left quadrant-stresses the functionality and complementarity of different governance processes. Informal governance is mainly a response to problems that cannot be effectively addressed in hierarchical political processes. In security policy, this is especially due to the increasing complexity and networking of threats and risks. 25 Informal deliberation and information-driven coordination are seen as superior to polarized political decisions. Various theoretical variants from EU studies and International relations, such as the literature on "regulatory networks", 26 "politics of expertise", 27 "experimental governance" 28 or "epistemic communities", 29 take a similar view. These approaches can also raise critical questions about the appropriate weight of technical expertise in political decision-making. Overall, however, flexible governance is seen as a tool for problem solving rather than political empowerment. 4. In the fourth perspective-related to the bottom-left quadrant-power relations are seen as fundamental driver of knowledge and expertise. In contrast to the second perspective, it is not so much the power of specific decision-making actors, as it is about the power of shaping broader definitions and problem definitions. Formal institutions and decision-making remain in place-or are not bypassedand are by-and-large effective, but they are shadowed, or partly overshadowed, by informal governance structures. In critical security research this comes in different variants, be it in the form of an integrated and competitive social field (à la Bourdieu) or as more separate professional networks and communities of practice. 30 In contrast to these deeper forms or patterns of informal governance, official decisions are either simply absent or strongly influenced in advance. A less explored variant of this argument about the power of professional security communities is the assumption of a more technologically structured power field dominated by the security industry. 31 24 Kleine, "Informal governance." 25 Nance and Cottrell [59]. 26 Maggetti and Gilardi [50]. 27 Liberatore [48]. 28 Sabel and Zeitlin [65], Monar [57]. 29 Cross [16]. 30 Pouliot [63], Scherrer and Dupont, "Noeuds ou champs?", Wesseling [79], Sheptycki, "Accountability." 31 Hayes et al. [34].
To be clear, this chapter cannot provide a systematic review or genuine competitive theory test for all four perspectives, but rather seeks to set out a wider agenda by relating emerging empirical patterns on European internal security governance to all of these different perspectives. To make this task more manageable, some additional preliminary remarks can be made. The first perspective on the replacement of informal by formal governance has frequently been a product of the general European integration process. The EU treaty competence for the creation of the so-called Area of Freedom, Security and Justice has built on previous informal intergovernmental arrangements, such as the TREVI group among European interior ministers. Similarly, the European Border and Coast guard was first based on an informal forum (the "External Border Practitioners Common Unit"). Yet, it is clearly not enough to limit the analysis to this conventional account of European integration and the complete displacement of informal governance by formal institutions.
The second perspective on the empowerment of executive actors through selective use of different governance venues has already been the subject of a detailed scientific debate. 32 By now, a strong reading that sees venue-shopping as a core driver of European security cooperation no longer seems valid. The catch-up of EU control mechanisms, such as the European Court of Justice, and the expansion of positive standards that restrict security professionals-whether asylum standards, data protection or criminal prosecutions-are evident. Yet, venue-shopping in the EU neighborhood, or in all external relations with third countries where appropriate rights and supervisory mechanisms cannot be asserted, 33 remains pertinent. Meanwhile, the specific role of great member states as the driver of informal governance within the EU needs more detailed evidence. A concrete example is the speculative influence of the G6 Ministers of the Interior 34 on the wider formation of the EU's internal security policy. Overall, executive empowerment cannot be seen as the master explanation, but remains relevant to European internal security governance. The following heuristic indicators on empirical governance patterns would buttress this presumption: (1) Informal forums and committees deal with sufficiently substantial tasks or policy results in order to work as circumvention of official decision-making structures; (2) actors who use these networks and forums are reasonably central and powerful; (3) informal networks for internal security policy are not shadowed by the corresponding networks for political or judicial oversight.
The third perspective on the functional role of informal governance is comparably underdeveloped in EU internal security. One can point to an emerging literature on the "politics of expertise," 35 but this remained based on a limited number of cases. Following some pioneering work with regard to epistemic communities, 36 32 Ette et al. [22], Kaunert and Léonard, "The development." 33 Lavenex, "Shifting up and out"; Menjívar [54]. 34 House of Lords [37], Die Linke [27]. 35 Boswell [7], Scholten and Verbeek [67], Mégie [53]. 36 Cross [17]. comparative research could develop on numerous networks for internal security coordination and probe the respective role of knowledge and expertise.
The fourth perspective of a transnational security field-or corresponding professional "guilds" 37 -should, in contrast, lead to networks by professional representatives rather than independent experts. Alternatively, the participation of private stakeholders would speak for the notion of concurrent economic interests of a new security industry. Overall, networks would not necessarily be linked to specific institutional structures, as professional field should develop relatively autonomously from other political processes. However, in the context of the following empirical overview, it must be admitted that this sociological perspective may not be fairly represented due to a lack of more detailed qualitative evidence.

Toward a broader understanding of European governance on internal security
The starting point for the empirical analysis is a database complied by the author, which currently contains 182 entries of different groups, forums and networks in different areas of European security governance. The original interest in the creation of this database was a simple cataloging, since an overall overview of these networks and groups is not available to date. On the official side, the EU provides a "Register of expert groups," which mainly works on policy preparation under the guidance of the European Commission. Only four, though comparatively large, networks are listed on the web site of the General Directorate of Internal Affairs. Forums and networks linked to the working groups of the Council Secretariat are only partly accessible. For example, one can refer to the "Law Enforcement Working Party," 38 which is supposed to exercise oversight over some networks among specialized police services. For other areas of internal security, there is no corresponding coordination point. However, regular small-scale inquiries from Bundestag delegates, in particular the left-wing faction, 39 as well as reports from Interior Ministries, 40 provide further insight into the range of governance activities. Finally, some cases have been derived from think tank studies, conference reports and occasional policy documents. In light of this fragmented picture, it is important to consolidate and validate comparative data on internal security governance. Therefore, the following discussion can only serve as a first cut at tracing comparative patterns of European security governance.

Networks across different areas of internal security
The first observation is the sheer number of transnational networks and forums, which deal with topics of police cooperation. This includes more than 50 groups, 38 Council of the European Union [14]. 39 Die Linke [25,26]. 40 Innenministerium Baden-Württemberg [38]. 37 Bigo [6].
while the areas of criminal cooperation and border security cover only 23 groups. Other areas of internal security, as identified by the EU, i.e., migration questions, cybersecurity and disaster control, are comparatively limited and are roughly equal to forums that address international counterterrorism (approx. 13-15 groups).
The amount of forums for police cooperation can be seen as an indicator of the breadth and differentiation of modern police tasks that go beyond the original areas of transnational "high policing." 41 For example, networks such as AIRPOL or the European medical and psychological experts' network for law enforcement (EMPEN) could be mentioned. There is a clear trend toward a reduction in the number of informal groups under the shadow of formal EU integration. Similar dynamics apply to the field of border security, where one can point to more specialized networks, such as the Frontex Risk Analysis Networks. Yet, this could be described as a controlled spread of networks for functional need-in contrast to the rather uncoordinated historical growth of forums for police cooperation. The area of criminal cooperation falls between these poles. On the one hand, new expert networks can be identified to support EUROJUST ("European Judicial Network"). On the other hand, cross-border cooperation has developed well outside of EUROJUST and launched decades before in the Council of Europe.
International counterterrorism is a particular area where the gap between publicly known information and actual pattern of international cooperation is presumed to be particularly large. 42 However, the number of known informal coordination forums is relatively limited, at least when compared to the rapidly growing EU legal framework for counterterrorism. For example, several decades-old forums, such as the "Police Working Group of Terrorism" or the "Club of Berne/Counterterrorism Group," continue to exist and only developed limited connections to the EU. 43 More dynamic and more recent developments have taken place in the area of radicalization. Since 2011, the European Commission has sought to promote an EU-wide network, which includes local authorities and street-level practitioners ("Radicalization Awareness Network") and attracts increasing amounts of official functions. This initiative complements further Member States-led informal groups ("Policy Planners Network") 44 and initiatives for the development of counter-narratives ("Syria Strategic Communication Network").
Overall, therefore, neither a clear logic of executive empowerment-which would lead to a stronger homogenization over topics-nor a general reduction in informal forums is born out in practice. Functional differentiation of subfields of EU internal security, such as the police work, as well as the interplay of EU integration and the related establishment of new expert groups seems to drive the number of observable governance forums. At the same time, the historical genesis of the security cooperation and issue-specific sensitivities, such as in the fight against terrorism, are also of lasting relevance. Thus, overall, it appears plausible that professional and 42 Hillebrand [36]. 43 Bures [9]. 44 Malmström [51]. sociological dynamics (perspective four) compete and overlap with strong functional growth of governance networks (perspective three).

Different network tasks
Next, the observed networks may be differentiated by inherent tasks. At a high level of abstraction, the most parsimonious typology has been provided by the original International Relations literature on transgovernmentalism. Slaughter distinguishes three tasks of transgovernmental networks across all possible issue topics, namely 1. information exchange; 2. regulatory tasks; and 3. executive tasks. 45 The preliminary empirical data collected on networks in EU internal security governance justify the following estimate: Approximately 75-85% of all recorded networks appear limited to mutual information exchanges (function 1). On the one hand, this may be explained by a relatively high number of expert groups in EU policy-making processes included in the database (currently 34). On the other hand, there are various professional groups in problem areas, which are not subject to uniform European regulation-such as dealing with drug-related criminality, the prevention of crime or "radicalization." As far as could be established on the basis of public information, the exchange of "best practices" and related soft coordination is the raison d'être of these networks.
Networks with regulatory tasks, as they are known from many economic areas, are not easily identifiable in the area of EU internal security. For example, international governance for the coordination of migration policy could, until now, only achieve limited convergence on lists of safe countries of origin. 46 However, only more detailed case studies can reveal how quasi-regulatory decisions can be made in networks that have no official regulatory tasks and responsibilities. The so-called ENLETS Group provides a pertinent example that has already been raised in various parliamentary questions. Officially, it is intended to provide information on possible technical innovations and research on internal security. The ENLETS Group could, however, also serve as an informal standard-setter for surveillance technology. This suspicion is warranted to the dynamics in technical standards, where first-timemover advantages may set important path dependencies. Similar questions about the creeping transition between information exchange and regulation arise in areas of border security. In addition to the official legal provisions on Schengen, new standards are being introduced via informal channels. This is currently the case with socalled foreign fighters, which is a term that is increasingly incorporated in criminal and administrative law. However, the first exchanges on this emerging security threat took place in informal working groups among national security authorities and later included EUROPOL 47 and other international organizations (INTERPOL, UN).
perspective on the importance of knowledge-driven interactions. The fourth sociological perspective is compatible with a significant emphasis on information exchange, insofar as it includes definition struggles and competitive agenda-setting of various professional groups. In contrast, executive networks would speak most strongly for the second perspective of executive empowerment. From a numerical point of view, there appears to be only a relatively limited number of such networks with operational responsibilities or tasks (current estimate 17). However, these cases are particularly sensitive. On the one hand, information on individual cases or ongoing investigations may be exchanged via such networks without having to take into account some national data protection or criminal law provisions. 48 On the other hand, some networks are designed for operational crisis management, whereby the executive powers remain as flexible as possible (see, for example, the Network of European Specialists "ATLAS" for counterterrorism invention teams).

Composition of the networks
The composition of networks, in so far as publicly accessible, provides further important insights into patterns of European internal security governance. Firstly, there is little evidence for the involvement of private stakeholders or private security actors, which contrasts with the wider academic debate on contemporary security governance. 49 The critique of an overrepresentation of commercial interests, 50 which is often criticized in other EU policies, does not apply at least to expert groups in this policy area. The only traceable structural connections with private companies are specific regulatory issues, such as the control of explosives and synthetic drugs, as well as cybersecurity where public-private partnerships are essential. 51 However, it should also be noted that the participation of other civil society groups appears limited. In addition to regular online consultations on specific policy proposals by the European Commission, which are often viewed critically as a mere legitimation strategy, 52 there is no framework for systematic relations with practitioners and civil society advocates. 53 This gap also explains the proposal to set up a European consultation forum on European security, which was introduced in April 2015. 54 Internationally leading non-governmental organizations, such as Amnesty International, are well-organized in Brussels and are interlinked with each other with respect to the EU's migration and asylum policy (ECRE). Moreover, the specialized NGO Statewatch provides a central reference point for all critical observers of the AFSJ. Looking at the wider field of EU internal security, however, there 51 Tropina and Callanan [74]. 52 Rasmussen and Carroll [64]. 53 Uçarer [75]. 54 European Commission [23, p. 3]. 48 Bayer [3], Alain [1]. 49 Dupont [19], Frevel and Schulze, "Kooperative Sicherheitspolitik." 50 Chalmers [10].
are only a few platforms and networks with civil society involvement, such as drug dependency or trafficking in human beings.
This means that public actors, whether they are representatives of security authorities, ministries or government agencies, remain central to European governance networks. This does not obviate further questions about membership and composition. On the one hand, politically oriented actors can compete with professional groups in such transgovernmental networks, as is accentuated differently in the perspectives for executive empowerment (2), functional governance (3) and professional fields (4). On the other hand, it is unclear to what extent the surveyed networks cover the EU as a whole, or to what extent subregional and issue-specific dynamics drive network membership. A variable geometry can be derived from functional considerations (range of certain threats, neighborly cooperation) as well as power policy components, if certain groups of states move forward within the European framework. Flexible integration and cooperation has been central historical feature of the AFSJ, 55 but also in more informal governance forums. Of the 182 cases in the database, at least a quarter (46) includes non-EU countries. In addition to the EFTA Schengen countries (N, CH, IS, LIE), the area of border security and migration control contains various networks and forums with both countries from the EU neighborhood as well as other Western countries (e.g., "Intergovernmental Consultations on Migration, Asylum and Refugees" or the "Budapest Process"). In police and criminal law formats, the Council of Europe, with its currently 47 member states, remains a reference point ("Groupe Pompidou," or the "European Cooperation Group on Undercover Activities"). Finally, the US and other OECD countries are emerging in the fields of financial control, drugs and the fight against terrorism ("Camden Asset Recovery Inter-Agency Network," "Dublin Group" or "International Working Group on Undercover Policing"). All in all, one can see both functional criteria for flexible governance beyond the EU, such as the integration of Schengen states or neighborhood states in border security issues, as well as approaches for "venue shopping" or executive coordination for other political reasons, as in differentiated networks of police and counterterrorism cooperation. The whole extent to which governance networks for internal security spread beyond the EU is difficult to determine and is presumably the greatest gap in the empirical assessment based on publicly available information.
At the same time, internal differentiation within the EU remains an important feature of internal security governance. Seventeen of the networks registered in the database contain less than all EU Member States. Germany, the Netherlands and the UK, however, are almost always represented. France, Belgium, Sweden, Spain and Italy follow closely behind in this relative ranking. The motives for the aboveaverage centrality of the three first-mentioned states are likely to differ, in particular between the relatively strong integration interest of Germany and the opposite pole of Great Britain, which is generally assumed to favor non-binding forms of cooperation. The Netherlands, in turn, long developed a special niche interest and pioneering role in the area of police cooperation, as underlined by Europol's headquarters and the most recent decision to provide a permanent home to the Counterterrorism Group of European intelligence services in The Hague. One may investigate further-compared to analyses conducted on the CSDP 56 -whether stable mini-lateral constellations among certain EU states can be observed in the area of internal security (see also above on the so-called G6 of Interior Ministers). This would substantiate the arguments about power-and actor-driven incentives for information governance that shadow or sidestep formal decision making (perspective 2).

Control and accountability
As mentioned in the introduction, many academic analysts express their distrust toward informal governance in the area of EU internal security. It is, however, noteworthy that there are hardly any empirical works that document specific control deficits. As a simple but very important criterion, one can start by searching for legal bases or publicly available administrative acts that underpin individual networks for internal security cooperation.
According to the data compiled by the author until today, it is estimated that up to a quarter of all networks and groups (182) appear to have such a (minimal) legal or official administrative basis. This is particularly the case for networks linked to EU institutions, such as the Commission, or agencies such as EUROPOL or FRON-TEX. However, it should be borne in mind that this basis can be very weak, such as a general decision by the Commission or the European Interior Ministers to set up a group. Such documents cannot, in themselves, provide an adequate legitimation of otherwise informal networks. Nevertheless, it should be noted that many networks for European security cooperation are not fundamentally beyond official decisionmaking processes, but are in some cases explicitly designed to play a complementary role.
In fact, a considerable number of governance networks are integrated with, or regularly supported by, EU agencies and institutions. This may allow for some indirect accountability. In EU studies, for example, the respective relationship and responsibility of the EU Commission for associated expert groups have been explored. 57 In particular, the European Parliament exerted pressure to render the composition of expert groups more transparent across all EU policy fields. Another set of networks (22) can be linked to EU agencies. Some of these networks, such as the Frontex Risk Analysis Network, are mentioned in annual activity reports and are thus subject to indirect external review. A better alternative is regular meeting reports as exemplified by the "European Network of contact points responsible for genocide, crimes against humanity and war crimes" that operated under the leadership of EURO-JUST. Overall, however, EU agencies have substantial room to improve the transparency of associated networks. This applies, for example, to the Europol Platform for Experts or the so-called Cross-Border Surveillance Working Group. In areas of 56 Wagnsson [78]. 57 Metz [55], Gornitzka and Sverdrup [31]. operational cooperation, the possible control deficit is particularly serious, even if the confidentiality of safety measures must be ensured.
A third group of networks is associated with working groups in the Council Secretariat, in particular with the so-called Law Enforcement Working Party. In the last few years, this working group has been responsible for a rudimentary coordination of different police networks. 58 These administrative documents, however, reveal very little about the work of different networks. A stronger model of accountability could be achieved in a visible reporting beyond expert committees or by publishing annual work programs. So far, this is only the case for some large networks financed by the Commission, such as the European Crime Prevention Network. Overall, internal security networks linked to different EU institutions offer some inroads to ensuring political accountability and matching complementarity with official EU policy objectives. Thus, a strategy of empowering or bypassing official channels (perspective 2) clearly does not apply in many instances. Yet, even in cases of EU-related networks, there is no uniform model of reporting, transparency or supervision.
In all other cases, it is even more difficult to identify a possible coordination center or rudimentary legal basis for political accountability, while language concerns among 28 member states complicate the research task. Non-EU-based networks are also particularly wide spread in areas of sensitive security cooperation, especially in the fields of police cooperation and counterterrorism (see above). One does not necessarily have to assume that professional autonomy (perspective 4) or political freedom of action of the network participants is maximized (perspective 2). A qualitative case study on the "Police Chiefs Task Force," which has now been resolved, came to the conclusion that the representatives of national police officers were closely controlled by national ministries of the interiowr. 59 However, the concern remains that these forms of transgovernmental networking offer incentives for a security practice that would not be enforceable in official structures. Thus, regular small queries [25,26] to such groups, such as, for example, in the case of the International Working Group on Undercover Policing, continue to be of high relevance.

Conclusion
This paper provides an initial empirical and conceptual mapping to work toward a more nuanced understanding of internal security governance in Europe. Firstly, it was argued that the interplay between formal EU integration steps and more informal European networks can take various competitive and complementary forms. Unfortunately, the existing literature on European internal security remains divided in distinct theoretical camps that uphold polarized views on the role, utility and legitimacy of these governance arrangements. To simplify, informal security governance networks either reflect positive professional or expertise-driven dynamics or are an illegitimate outgrowth of executive power and professional dynamics among security practitioners. We urgently need to go beyond these two camps and arrive at richer typologies and targeted critiques of the wide variety of governance arrangement for internal security in Europe.
Beyond condensing the different theoretical perspectives and bringing them into a dialogue with each other, there is a clear need to generate more data on the composition and tasks of observable governance networks. It is unsatisfactory to use individual case studies to extrapolate on wider dynamics of epistemic communities among security experts or executive empowerment among high-ranking officials. The potential impact of private actors as participants to internal security governance networks also needs further attention, not least as this area is likely to grow in the fields of cybersecurity and policing. It has also briefly been argued that the classic literature on transgovernmentalism deserves renewed attention in order to categorize governance networks on the basis of their intended function for information exchange, regulation or executive tasks. Clearly, there are different normative and practical challenges attached to these three different ideal types, even if, in practice, it may often be difficult to assign a particular governance network to only one or the other functionalist category. In short, it needs further critical investigation if we are, as appears on the basis of the present mapping exercise, mostly faced with information exchanges, whereas executive or regulative coordination remains rare in informal governance networks for European internal security. This question also applies to transatlantic or other geographic formats of cooperation, where it is typically claimed that governance participants mostly engage in an open-end exchange of views or best practices. More compulsory and hidden forms of security policy-making and operational coordination in transnational governance networks remain possible, however, and deserve focused attention. A well-known example is the Financial Action Task Force and related activities in the international fight against terrorism financing. But there are further comparable networks in transport security, technological security and possibly also surveillance, intelligence and policing.
Yet, overall, this paper suggests a cautious reading of the importance and relevance of informal governance arrangements for internal security. The most dominant pattern so far is the high degree of differentiation within and across different subfields of internal security, which may include some preferred venues for executive coordination, but overall remains within limited noncommittal exchanges of experiences and information. The next analytical step would be to provide a set of systematic case studies on the intersection or divergence of informal discussions in these networks with formal decision-making processes by the EU, for instance, on key decisions on the future of border security or transnational police cooperation. Furthermore, only qualitative research can highlight in how far network participants remain ultimately under the effective control and oversight of their home institutions and line ministries. These investigations could help to answer the overriding question, whether it is desirable and necessary, from the perspective of liberal democracy and the rule of law, to advance with further supranational institutionalization and legal regulation of internal security at the EU level, or whether the observable diversity and depth of informal governance networks already provide sufficient as well as legitimate security policy outcomes.